bug with non-standard evaluation and callr

[rmflight]

When using cc2 as part of an Rmarkdown report that is being managed by {callr} or a {drake} or {targets} workflow, and using a variable to hold a cutoff for significance, you may see the error that variable is not found (see reprex repo here).

This seems due to a less than ideal implementation of non-standard-evaluation on the part of the developer.
Because the significance call is really a filter operation, I think we should just depend on {dplyr::filter}, and pass the ... all the way down.
However, this means changing the statistical results to a proper data.frame, and other things.

To test the fix before mangling code across the statistical results class, etc, I think we should create a recursive function that upon reaching level 5 does the filtering and returns it, and verify it works via {callr}.

[rmflight]

I think the solution for this is the use of rlang::quosure() to make sure the value of the thing is captured, and kept with the thing being filtered.